TR2006-005

Worse is Better for Ambient Sensing



The intrinsic value of information coupled with the dramatically falling costs of networked sensors suggest that ambient intelligence and ubiquitous computing are inevitable. However, before society resigns itself to a world of constant observation and tracking, a process of moralization and ethical deliberation should occur. In this paper we examine the ethical implications of choosing camera networks or infrared motion detector networks. We employ the Dimensional Metaethics approach to help us structure examination of the complex issues involved. The analysis indicates that choice of sensor technology can powerfully affect the ethical landscape surrounding the final system. This paper also analyzes emprical results from questionnaires that asked participants to rate and choose between scenarios involving pan-tilt-zoom cameras and infrared sensors. In testing against a hypothetical even split in opinion, we find instead a significant preference for the scenario involving infrared sensors (p=0.007). The results show that significant proportion (73%, p=0.05) preferred a scenario with infrared sensors when compared to pan-tilt-zoom cameras. Participants also report that the scenario with infrared sensors was significantly less invasive and expressed a significantly weaker preference toward situations "Without Sensors that Collect Information About Location" (when compared with the scenario involving pan-tilt-zoom cameras). In short, we find that both dimensional metaethics and questionnaire results suggest that infrared sensors are better.